DEBUG saga convergence with sample_weight #31702

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Draft

ogrisel wants to merge 1 commit into scikit-learn:main from ogrisel:logistic-regression-sample-weight-saga-convergence

+21 −22

Member

ogrisel commented Jul 4, 2025

The purpose of this draft PR is to empirically show that our existing common test can reveal the convergence problem of the SAG and SAGA solver of logistic regression when fitting with large enough sample_weight values and low enough regularization.

At this stage, I do not recommend reviewing this PR with the intent to merge it, but rather use it as a concrete example to show that we should maybe change the common test to fail when the models do not converge to be able to better reveal sample_weight handling bugs.


          WIP saga convergence with sample_weight

0bf942c

github-actions bot added the module:utils label

github-actions bot commented Jul 4, 2025

✔️ Linting Passed

All linting checks passed. Your pull request is in excellent shape! ☀️

_{Generated for commit: 0bf942c. Link to the linter CI: here}

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels